In Reply to: MNI -> tlrc posted by Ed Vessel on November 14, 2002 at 19:30:09:

Hi all,

We are also registering our datasets to the average 152_T1 (using flirt) and then bringing them back into AFNI. Currently we have a script which brings these analyze datasets back into AFNI, in MNI space.

First, we have a 1x1x1 mprage anatomy, converted to analyze, and registered with flirt:
1)registering T1 (mprage) TO MNI avg152T1, output a matrix
2)applying matrix to T1, keep FOV of MNIavg152 space, but 1x1x1 resolution(
use - applyisoxfm 1)

3) Here is the to3d command line to bring this MNI registered dataset back into AFNI:
to3d -prefix "subj"_T1_MNI_reg -session "path" -anat -view tlrc -xFOV
90L-92R - yFOV 126P-92A -zSLAB 71.5I-109.5S -orient LPI "path"
$subj"_T1_MNI_reg.hdr

Analyze volumes in MNI space are oriented as LPI, and thus need to be read into AFNI that way. The offsets above are set to put a 1x1x1 dataset that has been registered to the 152_T1 into AFNI, the origin set at the MNI origin of 90L 126P 72I

I derived this from the MNI colin27 dataset (one subject anatomy, MNI 1x1x1, scanned 27 times, the origin is at 90L 126P 72I. This is where our to3d command differs slightly from the one that Ed Vessel mentions.

The FOVs for MNI space appear to be: 182x, 218y, 182z

We also register our functional echo-planar datasets to the subject's MNI registered T1 and bring that into AFNI as well. We are setting the view to tlrc so that we may use the volume renderer with our datasets. We are currently exploring options for automatically reporting Brodmann areas on these datasets.

Keep in mind that we are still playing around with this, but so far, it is fairly automated and seems to work fine. I'd be interested in what others think about this issue, or if anyone else has another method of doing this.

Tim Souza